GRASP: Grammar- and Syntax-based Pattern-Finder for Collocation and Phrase Learning

GRASP: Grammar- and Syntax-based Pattern-Finder for Collocation and Phrase Learning Mei-hua Chen a, Chung-chi Huang a, Shih-ting Huang b, and Jason S. Chang b aInstitute of Information Systems and Applications, National Tsing Hua University, HsinChu, Taiwan 300, R.O.C. {chen.meihua, u901571}@gmail.com bDepartment of Computer Science, National Tsing Hua University, HsinChu, Taiwan 300, R.O.C. {koromiko1104, jason.jschang}@gmail.com Abstract. We introduce a method for learning to find the representative syntax-based context of a given collocation/phrase. In our approach, grammatical patterns are extracted for query terms aimed at accelerating lexicographers’ and language learners’ navigation through the word usage and learning process. The method involves automatically lemmatizing, part-of-speech tagging and shallowly parsing the sentences of a large-sized general corpus, and automatically constructing inverted files for quick search. At run-time, contextual grammar patterns are retrieved and presented to users with their corresponding statistical analyses. We present a prototype system, GRASP (grammar- and syntax-based pattern-finder), that applies the method to computer-assisted language learning. Preliminary results show that the extracted patterns not only resemble phrases in grammar books (e.g., make up one’s mind) but help to assist the process of language learning and sentence composition/translation. Keywords: Computer-assisted language learning, collocation, part-of-speech tagging, grammatical patterns, and inverted files. 1 Introduction Many language learners’ queries (e.g., “play” and “role”) are submitted to language-learning tools on the Web every day and an increasing number of services on the Web specifically target second language learning. For example, Word Sketch Engine (www.sketchengine.co.uk) is a concordancer that automatically summarizes a word’s grammatical and collocational behavior while services such as TANGO1 and MUST2 provide a means of collocation finding (e.g., verb- noun and adjective-noun collocations) and collocation correcting. Language-learning tools such as Word Sketch and TANGO typically accept only one querying word and retrieve sentences with it or words it co-occurring probabilistically more frequently than usual. However, learners may attempt to learn the usage of a certain word sense of the query word. Consider the polysemy “play”. In WordNet, it has a myriad of senses including ‘participating in games or sport’, ‘act or having an effect in a specified way’ and ‘play on an instrument’. Learners may be aware of its senses but intend to acquire more knowledge on the context or usage of the word sense ‘act or having an effect in a specified way’. Suggested by (Yarowsky, 1995), accompanying “play” with its collocate “role” is a good way to narrow down 1 candle.fl.nthu.edu.tw/collocation/webform2.aspx 2 candle.fl.nthu.edu.tw/vntango/ PACLIC 24 Proceedings 357 the senses of “play” (and vice versa). Therefore, multi-word query might be as important in language learning. However, the best response to the multi-word query “play role” is probably not an overwhelming set of sentences with it which may be returned by general-purpose concordancers and search engines (e.g., Google). A good response might indicate that the collocation “play role” is frequently followed by the grammatical part-of-speech (PoS) patterns ‘preposition determiner’ (e.g., “in the” and “in this”), ‘preposition noun’ (e.g., “in society” and “in relation”) and ‘preposition gerund’ (e.g., “in determining” and “in shaping”), and preceded …

Content truncated to 500 words